Skip to content

Pull requests: flagos-ai/FlagGems

New pull request New
669 Open 4,599 Closed
669 Open 4,599 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

[MTHREADS] Conv1D optimization investigation notes (no code change) documentation Improvements or additions to documentation size/Small
#6037 opened Sep 7, 2026 by gao0624 Contributor Loading…
test(flash_attention): support --ref cpu reference in backward accuracy test ops/aten size/Medium tests
#6036 opened Sep 7, 2026 by bin913 Collaborator Loading…
4 of 10 tasks
1
ci(aliases): map dispatch_fused_moe_kernel to test_fused_experts_impl size/XSmall
#6035 opened Sep 7, 2026 by bin913 Collaborator Loading…
3 of 8 tasks
1
test(clone): add clone accuracy tests so pytest -m clone collects cases size/Small tests
#6034 opened Sep 7, 2026 by bin913 Collaborator Loading…
3 of 9 tasks
1
test(and_scalar): tag bitwise_and_scalar tests so pytest -m and_scalar collects size/XSmall tests
#6033 opened Sep 7, 2026 by bin913 Collaborator Loading…
3 of 9 tasks
1
fix(cdist_backward): dispatch gradient formula by p-norm ops/aten size/Medium tests
#6032 opened Sep 7, 2026 by bin913 Collaborator Loading…
6 of 11 tasks
1
test(perf): add missing benchmark markers for the 27 operators reported in #4947 benchmark size/Large tests
#6031 opened Sep 7, 2026 by bin913 Collaborator Loading…
5 of 11 tasks
1
[QC][Ascend] Add W8A16 FP8 TopK operator benchmark core size/XLarge tests vendor/Ascend
#6030 opened Sep 7, 2026 by July-h5kf3 • Draft
update to triton3.8.0 core ops/aten ops/fused size/Medium
#6029 opened Sep 7, 2026 by huangyiqun Collaborator • Draft
3 tasks
1
[KMCompiler][Ascend] Add ascend backend for igammac benchmark size/XLarge tests vendor/Ascend
#6028 opened Sep 7, 2026 by LittleShun1214 Contributor Loading…
1 of 3 tasks
[MTHREADS] fix polar: correct output shape for N-D inputs size/Medium vendor/MooreThreads
#6026 opened Sep 6, 2026 by gao0624 Contributor Loading…
[MTHREADS] optimize unique2 post-sort pipeline size/Large vendor/MooreThreads
#6025 opened Sep 6, 2026 by gao0624 Contributor Loading…
fix(broadcast_tensors): handle zero-size dims in broadcast shape computation ops/aten size/Small
#6024 opened Sep 5, 2026 by cccxxttt Loading…
1
[MTHREADS] optimize quantile order-statistic selection size/Medium vendor/MooreThreads
#6022 opened Sep 5, 2026 by gao0624 Contributor Loading…
1 of 3 tasks
[KernelGen][Nvidia] Add put_ operator with Triton kernel benchmark core KernelGen ops/aten size/Medium tests
#6021 opened Sep 5, 2026 by Yukun-Cui Contributor Loading…
[KernelGen][Nvidia] Add put operator with Triton kernel benchmark core KernelGen ops/aten size/Large tests
#6020 opened Sep 5, 2026 by Yukun-Cui Contributor Loading…
[MTHREADS] Optimize avg_pool2d backward size/Medium vendor/MooreThreads
#6019 opened Sep 5, 2026 by gao0624 Contributor Loading…
perf(FLA): optimize fused recurrent KDA decode with Triton benchmark ops/fused size/XLarge tests
#6015 opened Sep 4, 2026 by huangyiqun Collaborator • Draft
1 of 3 tasks
[KMCompiler][Ascend] Opt dist operator on Ascend. benchmark size/Medium tests vendor/Ascend
#6014 opened Sep 4, 2026 by Caeruleann Collaborator Loading…
feat(pt2): add torch.compile support for FlagGems kernels used by vLLM 0.24 core ops/aten ops/fused size/XLarge vllm
#6013 opened Sep 4, 2026 by JiaryCoder Loading…
1 of 3 tasks
[KernelGen][Nvidia] Add row_indices operator with Triton kernel benchmark core KernelGen ops/aten size/Medium tests
#6012 opened Sep 4, 2026 by Yukun-Cui Contributor Loading…
[KernelGen][Nvidia] Add row_stack operator with Triton kernel benchmark core KernelGen ops/aten size/Medium tests
#6011 opened Sep 4, 2026 by Yukun-Cui Contributor Loading…
[kunlunxin] register new ops and update backend configs core size/Medium vendor/Kunlunxin
#6009 opened Sep 4, 2026 by RobertLuobo Contributor Loading…
3 tasks
[KMCompiler][Ascend] ascend masked_scatter_backward benchmark size/Large tests vendor/Ascend
#6008 opened Sep 4, 2026 by YangLong114514 Contributor Loading…
[KMCompiler][Ascend] Opt unsafe_index operator on Ascend. benchmark size/Large tests vendor/Ascend
#6007 opened Sep 4, 2026 by Caeruleann Collaborator Loading…
ProTip! Find all pull requests that aren't related to any open issues with -linked:issue.